Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.04488 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912630464053248 |
|---|---|
| author | Chang, Edward Y. Chang, Ethan Y. |
| author_facet | Chang, Edward Y. Chang, Ethan Y. |
| contents | Multi-agent debate often wastes compute by using a fixed adversarial stance, aggregating without deliberation, or stopping on heuristics. We introduce MACI, an active controller with two independent dials that decouple information from behavior: an information dial that gates evidence by quality, and a behavior dial that schedules contentiousness from exploration to consolidation. A moderator tracks disagreement, overlap, evidence quality, and argument quality, and halts when gains plateau. We provide theory-lite guarantees for nonincreasing dispersion and provable termination, with a budget-feasible scheduler. Across clinical diagnosis and news-bias tasks, MACI improves accuracy and calibration while reducing tokens, and converts residual uncertainty into precision RAG plans that specify what to retrieve next. We use a cross-family LLM judge (CRIT) as a conservative soft weight and stop signal, validated for order invariance and judge-swap stability; stability depends on using high-capability judges. MACI turns debate into a budget-aware, measurable, and provably terminating controller. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_04488 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Multi-Agent Collaborative Intelligence: Dual-Dial Control for Reliable LLM Reasoning Chang, Edward Y. Chang, Ethan Y. Artificial Intelligence Information Theory I.2.4 Multi-agent debate often wastes compute by using a fixed adversarial stance, aggregating without deliberation, or stopping on heuristics. We introduce MACI, an active controller with two independent dials that decouple information from behavior: an information dial that gates evidence by quality, and a behavior dial that schedules contentiousness from exploration to consolidation. A moderator tracks disagreement, overlap, evidence quality, and argument quality, and halts when gains plateau. We provide theory-lite guarantees for nonincreasing dispersion and provable termination, with a budget-feasible scheduler. Across clinical diagnosis and news-bias tasks, MACI improves accuracy and calibration while reducing tokens, and converts residual uncertainty into precision RAG plans that specify what to retrieve next. We use a cross-family LLM judge (CRIT) as a conservative soft weight and stop signal, validated for order invariance and judge-swap stability; stability depends on using high-capability judges. MACI turns debate into a budget-aware, measurable, and provably terminating controller. |
| title | Multi-Agent Collaborative Intelligence: Dual-Dial Control for Reliable LLM Reasoning |
| topic | Artificial Intelligence Information Theory I.2.4 |
| url | https://arxiv.org/abs/2510.04488 |